The Astrophysical Journal, 876:113 (9pp), 2019 May 10 https://doi.org/10.3847/1538-4357/ab16e4 © 2019. The American Astronomical Society. All rights reserved.
Stellar Rotation and the Extended Main-sequence Turnoff in the Open Cluster NGC 5822
Weijia Sun1 , Richard de Grijs2,3 , Licai Deng4,5,6 , and Michael D. Albrow7 1 Kavli Institute for Astronomy & Astrophysics and Department of Astronomy, Peking University, Yi He Yuan Lu 5, Hai Dian District, Beijing 100871, Peopleʼs Republic of China 2 Department of Physics and Astronomy, Macquarie University, Balaclava Road, Sydney, NSW 2109, Australia 3 International Space Science Institute—Beijing, 1 Nanertiao, Hai Dian District, Beijing 100190, Peopleʼs Republic of China 4 Key Laboratory for Optical Astronomy, National Astronomical Observatories, Chinese Academy of Sciences, 20A Datun Road, Chaoyang District, Beijing 100012, Peopleʼs Republic of China 5 School of Astronomy and Space Science, University of the Chinese Academy of Sciences, Huairou 101408, Peopleʼs Republic of China 6 Department of Astronomy, China West Normal University, Nanchong 637002, Peopleʼs Republic of China 7 School of Physical and Chemical Sciences, University of Canterbury, Private Bag 4800, Christchurch, New Zealand Received 2019 February 2; revised 2019 March 20; accepted 2019 April 6; published 2019 May 9
Abstract The origin of extended main-sequence turnoffs (eMSTOs) in intermediate-age (1–3 Gyr) clusters is one of the most intriguing questions in current star cluster research. Unlike the split main sequences found in some globular clusters, which are caused by bimodal populations in age and/or chemical abundances, eMSTOs are believed to be due to stellar rotation. We present a spectroscopic survey of MSTO stars in a nearby, intermediate-age (0.9 Gyr), 3 low-mass (∼1.7×10 Me) Galactic open cluster, NGC 5822. We derive a clean sample of member stars based on Gaia proper motions and parallaxes and confirm the existence of an eMSTO. Using medium-resolution (R∼4000) Southern African Large Telescope spectra, we derive the rotational velocities of 24 member stars (representing 20% completeness around the eMSTO region) and find that the loci of the main-sequence stars in the eMSTO region show a clear correlation with the projected rotational velocities in the sense that fast rotators are located on the red side of the eMSTO and slow rotators are found on the blue side. By comparison with a synthetic cluster model, we show that the stellar rotational velocities and the eMSTO of NGC 5822 can be well reproduced, and we conclude that stellar rotation is the main cause of the eMSTO in NGC 5822. Key words: galaxies: star clusters: general – open clusters and associations: individual (NGC 5822) – stars: rotation
1. Introduction Aided by Gaia Data Release 2 (DR2), we can derive “clean” The extended main-sequence turnoff (eMSTO) phenomenon samples of open clusters in the Milky Way free of contamina- tion by field stars (Cantat-Gaudin et al. 2018). The discovery of —i.e., the notion that the main-sequence turnoff (MSTO) in the eMSTOs in Galactic open clusters, similar to those observed in color–magnitude diagram (CMD) is much wider than the Magellanic Cloud clusters, has opened up a new chapter in our prediction from single stellar population modeling—first comprehension of the formation of eMSTOs and the evolution discovered in NGC 1846 by Mackey & Broby Nielsen of open clusters. Cordoni et al. (2018) found the existence of (2007), is a common feature found in a large fraction of eMSTOs in all 12 open clusters younger than ∼1.5 Gyr they young and intermediate-age (2 Gyr) massive Large and Small analyzed (including NGC 5822), suggesting that eMSTOs are a Magellanic Cloud clusters (e.g., Mackey et al. 2008; Milone common feature of intermediate-age clusters in the Milky Way et al. 2009, 2015; Goudfrooij et al. 2011; Li et al. 2014b; and that they are regulated by the same mechanism as that Correnti et al. 2017). operating in Magellanic Cloud clusters. The understanding of In the past few years, our understanding of the eMSTO open clusters, which used to be considered prototypes of a phenomenon in these clusters has been enriched and enhanced single stellar population, encountered a significant upheaval significantly. Rather than due to intrinsic age spreads, stellar due to the discovery of eMSTOs in Galactic open clusters. It rotation is believed to play an important role in shaping the has been reported that stars exhibit a wide range of rotation morphology of the eMSTO (e.g., Bastian & de Mink 2009;Li rates both in the field (Huang et al. 2010) and in open clusters et al. 2014a; Niederhofer et al. 2015). This theory is further (Huang & Gies 2006). A number of studies have explored the reinforced by multiple lines of photometric and spectroscopic effect of stellar rotation on the CMD around the MSTO region evidence. Using narrowband photometry, Bastian et al. (2017) (Brandt & Huang 2015a, 2015b, 2015c). However, the direct detected a large fraction (∼30%–60%) of Be stars in the MSTO connection between the stellar rotation rates of MSTO stars and regions of NGC 1850 (∼80 Myr) and NGC 1856 (∼280 Myr), their loci in the CMD was only revealed recently. Bastian et al. favoring the interpretation that their split main sequences are (2018) used Very Large Telescope/FLAMES spectroscopy of caused by the effects of fast rotators. Similar mechanisms were 60 cluster members in NGC 2818, an 800 Myr open cluster, to later confirmed in NGC 1866 (∼200 Myr) and NGC 1818 measure the stellar rotational velocities and found that stars (∼40 Myr) through high-resolution spectroscopic surveys, exhibiting high rotational velocities are located on the red side suggesting that these clusters host a blue main sequence of the eMSTO and those rotating slowly are on the blue side, in composed of slow rotators and a red one composed of fast agreement with the prediction of the stellar rotation scenario. rotators (Dupree et al. 2017; Marino et al. 2018b). Marino et al. (2018a) also reported that the multiple sequences
1 The Astrophysical Journal, 876:113 (9pp), 2019 May 10 Sun et al.
Table 1 Observation Log of the SALT Runs Na ( ) ( ) Mask Name Program aJ2000 dJ2000 Exp. Time s Date UT (1)(2)(3)(4)(5)(6)(7) NGC5822p2 2017-2-SCI-038 15h04m33 43 −54°26′33 16 13 600 2018 Feb 8 2018-1-SCI-006 15h04m33 43 −54°26′33 16 13 764 2018 Aug 3 NGC5822p3 2017-2-SCI-038 15h04m19 43 −54°16′44 52 10 600 2018 Apr 26 NGC5822p4 2017-2-SCI-038 15h03m11 44 −54°16′22 93 11 600 2018 Feb 8 2018-1-SCI-006 15h03m11 44 −54°16′22 93 11 764 2018 Jul 30 NGC5822p6 2017-2-SCI-038 15h03m09 95 −54°31′46 39 9 600 2018 Feb 11 2018-1-SCI-006 15h03m09 95 −54°31′46 39 9 764 2018 Aug 14 NGC5822p8 2017-2-SCI-038 15h05m10 13 −54°19′41 91 7 600 2018 Feb 26 NGC5822p9 2017-2-SCI-038 15h04m36 35 −54°34′51 48 5 600 2018 Apr 30
Note. a Number of science slits in each field. they found in the young cluster NGC 6705 correspond to stellar of ∼4000 with a per-pixel resolution of 0.33 Å at a central populations with different rotation rates. wavelength of 4884.4 Å. Regular bias, argon arc lamp, and In this paper, we present a spectroscopic survey of MSTO quartz lamp flat-field calibration frames were taken as part of stars in the nearby (∼760 pc) intermediate-age (0.9 Gyr) open normal SALT operations. We used the PySALT package cluster NGC 5822. We find the presence of an eMSTO in this (Crawford et al. 2010) to perform the primary reduction and cluster and verify that it is not an artifact caused by differential wavelength calibration. For all of our samples, we obtained extinction. The loci of the main-sequence stars in the eMSTO spectra with a signal-to-noise ratio (S/N) per pixel in excess region show a clear correlation with the projected rotational of 200. velocities, with fast rotators lying on the red side of the eMSTO and slow rotators on the blue side. By comparison with a 2.2. Membership Determination synthetic cluster within the framework of stellar rotation, we Gaia ( argue that the observed morphology of the eMSTO in the CMD We exploited the DR2 Gaia Collaboration et al. ) can be properly explained by the model and that stellar rotation 2016, 2018 to analyze the stellar photometry, proper motions, is likely the main contributor to the eMSTO morphology in and parallaxes, and to perform membership determination in fi NGC 5822. the NGC 5822 eld. First, we acquired the stellar catalog from Gaia ( ′ This article is organized as follows. In Section 2 we present the database within 2.5 times the cluster radius 35 ; Dias ) the observations, data reduction procedures, and membership et al. 2002 . In the vector-point diagram of stellar proper motions, NGC 5822 showed a clear concentration centered at determination. Section 3 reports our main results, showing a -1 strong correlation between the stellar rotation rates and their (mqmadcos ,)(»- 7.44, - 5.52 ) mas yr . The other overden- -1 loci in the CMD region covered by the eMSTO. A discussion sity located at (mqmadcos ,)(»- 3.67, - 2.52 ) mas yr cor- and our conclusions are summarized in Section 4. responds to a nearby cluster, NGC 5823. Then, we derived the 22 quantity mR =-áñ+-áñ()()mqmqaacos cos mm ddand -1 2. Data Reduction and Analysis applied a cut of mR = 0.4 mas yr to conduct our primary 2.1. Spectroscopic Data membership selection. Next, we placed a further constraint on the parallaxes by estimating the mean parallax of the proper- We selected spectroscopic candidates in NGC 5822 using motion-selected stars (ávñ=1.18 mas yr-1) and adopted stars the photometric survey in the UBVI and uvbyCaHb systems with parallaxes within 0.115 mas yr-1 as cluster members. undertaken by Carraro et al. (2011). These broadband Note that this approach is slightly different from that adopted observations were obtained with the Y4KCAM camera by Cordoni et al. (2018), in the sense that we adopted a straight ϖ mounted on the Cerro Tololo Inter-American Observatory cut in both mR and rather than applying different selection (CTIO) 1 m telescope and the intermediate- and narrowband criteria for stars of different brightnesses. One reason for this imaging was carried out using the CTIO 0.9 m telescope. approach is that NGC 5822 is sufficiently close that its member Through a cross-correlation with the UCAC3 database stars can be easily separated from field stars using parallaxes (Zacharias et al. 2000), these authors derived 136 probable (see the top right panel of Figure 1). On the other hand, the photometric members and 322 probable nonmembers. limited number of stars in NGC 5822 makes it hard to reliably We obtained spectroscopic observations with the Southern calculate the corresponding rms for each magnitude bin. As we African Large Telescope (SALT; Buckley et al. 2006) did not set out to compile a homogeneous database for multiple equipped with the Robert Stobie Spectrograph (RSS) using clusters, our approach is suitable for our analysis of this single its multi-object spectroscopy (MOS) capability over nine nights cluster. We present the spatial distribution as well as the CMD from 2018 February 8/9 to August 14/15 under programs of the member stars of NGC 5822, together with all stars in the 2017-2-SCI-038 and 2018-1-SCI-006. Six masks were field, in the bottom panels of Figure 1. We present the CMD of designed to cover 88 stars (including repetitions) in NGC NGC 5822 color-coded by the stellar classifications based on 5822 as part of program 2017-2-SCI-038, with three masks their loci in Figure 2. Member stars classified as MSTO, MS, observed for the second time the following semester (see and giant stars are marked as green squares, blue triangles, and Table 1). The PG2300 grating was used with a 1 arcsec wide red diamonds, respectively. Member stars with spectroscopic short slit binned 2×2, offering a nominal spectral resolution data are presented using solid markers and field stars with
2 The Astrophysical Journal, 876:113 (9pp), 2019 May 10 Sun et al.
Figure 1. (Top left) Vector-point diagram of the proper motions for stars brighter than G = 15 mag within 87 5 of the center of NGC 5822. The red circle shows the primary selection of cluster members. (Top right) G band vs. parallaxes. The primary members selected from the proper motions are marked as solid dots, and the parallax-selected members are marked as red dots. The vertical dashed lines represent the selection criteria applied to the parallaxes. (Bottom left) Spatial distribution of stars selected, with cluster members highlighted as red points. (Bottom right) CMD of all stars in the field (gray dots) and member stars of NGC 5822 (red solid dots). The eMSTO is visible around G ~ 11.5 mag. spectroscopic data are shown as gray circles. Following the degree of the spatial variation of the reddening and found decontamination of the field stars, 24 member stars (21 MSTO that its influence is negligible compared with the extent of the and 3 MS stars) were left in our observational sample; 13 eMSTO. Given its close distance and low Galactic latitude, we member stars were observed a second time. We estimate that found that we could not use a 2D reddening map (e.g., Schlafly the total number of member MSTO stars in this cluster is ∼107, & Finkbeiner 2011) to estimate the differential reddening. suggesting that the completeness of our observed sample is Instead, we adopted the method of Nataf et al. (2013), who around 20% in the eMSTO region. assumed a two-component model for the distribution of the This cluster shows a clear eMSTO feature around dust, including the mean density of dust along the plane r and ~ D G 11.5 mag. To further demonstrate that this is not an a scale height H . Therefore, the prediction for the reddening in artifact owing to residual differential reddening, we estimated D
3 The Astrophysical Journal, 876:113 (9pp), 2019 May 10 Sun et al.
with effective temperatures, Teff, ranging from 5000 K to 8000 K (in steps of 100 K), surface gravities from logg = 3.5 to logg = 5.0 (in steps of 0.1), and metallicities from [Fe/ H]=−1.0 to [Fe/H]=1.0 dex (in steps of 0.5 dex) from the Pollux database (Palacios et al. 2010). We applied the latest ATLAS12 model atmospheres (Kurucz 2005) where blanketed model atmospheres handle line opacity in stellar atmospheres using the Opacity Sampling technique. The models assume a plane parallel geometry, hydrostatic and radiative equilibrium, as well as local thermodynamic equilibrium. The microturbu- lent velocity was fixed to 2 km s-1 for all models. Synthetic spectra were then generated using the SYNSPEC tool (Hubeny & Lanz 1992). Each model spectrum was convolved with the rotational profile for a given rotational velocity and imple- mented with an instrumental broadening as well as a radial velocity shift. Given that the light enters through off-axis slits (in the dispersion direction) in the MOS, the actual resolution may vary from slit to slit and from mask to mask. Therefore, we adopted the full width at half maximum (FWHM) of the corresponding arc lines as an indicator of the instrumental broadening effect. Then, we used the Markov chain Monte Carlo (emcee; Foreman-Mackey et al. 2013) method to sample fi Figure 2. CMD of NGC 5822 color-coded by the stellar classi cations based the five-dimensional parameter space (visin , vr, Teff, log g, on their loci. Member stars classified as MSTO, MS, and giant stars are marked [Fe/H]) to employ a c2 minimization. For each of the 3000 as green squares, blue triangles, and red diamonds, respectively. Member stars 2 with spectroscopic data are presented using solid markers, and field stars with runs of the MCMC procedures, c values and their associated 2 spectroscopic data are shown as gray circles. probabilities e-c 2 were stored. Probability distributions were then generated by projecting the sum of the probabilities onto a given direction is given by the dimension considered. A Gaussian fit to the distribution provides its width σ, which we adopt as the uncertainty. d To estimate the influence of instrumental broadening on the EB()-= VrD exp ((∣∣)) - r sin b Hdr , () 1 ò0 determination of the rotational velocities, we generated a set of mock spectra by sampling the projected rotational velocities where b represents Galactic latitude and d is the distance visin from 20 to 200 km s-1, assuming a uniform S N= 200 derived from the corresponding Gaia parallax. The distance and a reasonable uncertainty for the instrumental broadening ( ) fi was derived from the parallax by implementing the formalism sFWHM = 0.1 Å , and we measured the best- tting parameters of Astraatmadja & Bailer-Jones (2016). H = 164 pc is the dust from those mock spectra. We repeated this procedure 100 times scale height, and r = 0.427 mag kpc-1 (Nataf et al. 2013). and estimated the median values and the 68th percentiles of the D velocity distribution. In Figure 4 we present a comparison of The average extinction for cluster members is around the rotational velocities of the mock data with those derived E ()BV-=0.126 mag with a standard deviation of through profile fitting. The blue shadowed region corresponds ( ) sEB()- V = 0.004 mag. Using the Cardelli et al. 1989 and to 1σ and the one-to-one relation is indicated by an orange solid O’Donnell (1994) extinction curve with RV=3.1, we line. Given the intermediate spectral resolution, it is hard to corrected the reddening to the average reddening value. In differentiate the effect of rotational from instrumental broad- Figure 3 we present the CMD of selected cluster member stars ening for slow rotators. Therefore, we defined the detection before (left) and after (right) our differential reddening limit of the rotational velocity to be the velocity where its correction. A visual inspection suggests that the morphology uncertainty is around half of the actual value and, for slow -1 of the CMD remains unchanged and the eMSTO still exists rotators with visin 55 km s , the uncertainties of the after having applied this correction. measurements are comparable to their actual values, while ’ the uncertainty is less than 5% and 3% for the mock spectra We used the Padova group s PARSEC 1.2S isochrones -1 -1 (Bressan et al. 2012) to perform our CMD fits based on visual with visin 100 km s and visin 150 km s , matching. Cordoni et al. (2018) derived an age of ~1Gyrand a respectively. solar-like metallicity (Z = 0.0152). Our best fit agrees with these results (Figure 3). The best-fitting isochrone has an age of 3. Extended MSTOs and Stellar Rotation 0.9 Gyr for Z = 0.017 and a distance of ~760 pc. The binary The eMSTO of NGC 5822, if interpreted as an age sequence is clearly visible in the CMD and Cordoni et al. difference, is around 300– 350 Myr. Cordoni et al. (2018) (2018) estimated the fraction of unresolved binaries with estimated the ages of the stars around the eMSTO region by q > 0.7 at 0.131. linearly interpolating a grid of isochrones and calculated the ’ 2.3. Rotational Velocities FWHM of the cluster s age distribution, which gives a spread of 270 52 Myr. They also showed that the FWHM of the The projected rotational velocities were measured by fitting NGC 5822 eMSTO follows the correlation between the width the absorption line profiles of Hβ and the Mg I triplet. We of the eMSTO and cluster age applicable within the framework compiled a library of high-resolution synthetic stellar spectra of stellar rotation.
4 The Astrophysical Journal, 876:113 (9pp), 2019 May 10 Sun et al.
Figure 3. Comparison of the CMD of selected cluster member stars before (left) and after (right) differential reddening correction. Different colors represent different extinctions E (BV- ). The best-fitting isochrone for the bulk stellar population is shown in the right panel as a red solid line. The best-fitting isochrone (red solid line) has an age of 0.9 Gyr with Z = 0.017 and a distance modulus of around 9.4 (~760 pc). Isochrones for ages of log()t yr-1 = 9.05 and 9.15 are also overplotted (red dashed lines).
We also compared the observed cluster data with a synthetic cluster data set that included the effects of stellar rotation. The synthetic cluster data were derived from the SYCLIST models (Georgy et al. 2013, 2014), assuming a metallicity of Z = 0.014, an age of log()t yr-1 = 8.95, and a binary fraction of 0.131, with a rotational distribution derived from Huang et al. (2010) and a random rotation axis distribution. The model also accounts for the limb-darkening effect (Claret 2000) as well as for the gravity darkening law of Espinosa Lara & Rieutord (2011). In the left panel of Figure 6, the synthetic cluster is superposed onto the CMD of NGC 5822 and the eMSTO feature is well reproduced and consistent with coeval stellar populations with different rotation rates. The projected rotational velocity of the synthetic cluster follows a similar trend as the real member stars, which become redder as the stellar rotation rates increase. In the middle panel we present a realistic synthetic cluster with a number of stars comparable to that in the observed CMD. To provide a better comparison with the simulation, we introduced the pseudo-color D()GGbp- rp as the normalized color difference with respect to the blue ridgeline in the Figure 4. Rotational velocities of the mock data (horizontal axis) vs. rotational direction determining how stellar rotation may change the locus velocities derived from profile fitting. The blue shadowed region corresponds of a star in the CMD (black arrow) to represent the deviation in to 1σ and the one-to-one relation is indicated by an orange solid line. The red dashed lines represent the lower limit of reliable measurements of the rotational color that may be caused by stellar rotation. We adopted the velocity (55kms-1) where the uncertainty is around half of the actual value. blue edge of the synthetic cluster, which represents the population of nonrotating stars, as the fiducial ridgeline. In In Figure 5, we present the CMD of NGC 5822, with the the right panel of Figure 6, the D()GGbp- rp versus visin member stars color-coded by their rotational velocities. We diagram for all stars with projected rotational velocity found that their loci in the CMD region covered by the eMSTO measurements is shown, and the gray dots represent the same strongly depend on stellar rotation, in the sense that rapid distribution for the synthetic cluster. We found that most of our rotators tend to lie on the red side of the eMSTO while slow targets follow the trend predicted by the stellar rotational rotators are usually found on the blue side. Similar results have model, where the pseudo-color is close to zero for slow rotators also been discovered in young and intermediate-age clusters in and it increases significantly as the rotational velocity the Magellanic Clouds (Dupree et al. 2017; Kamann et al. increases. Two outliers in the right panel of Figure 6 (Gaia 2018), as well as in Galactic open clusters (Bastian et al. 2018; IDs: 5887669198096568960 and 5887671397119565312), Marino et al. 2018a). The stellar structural parameters as well which have relatively large pseudo-colors compared with their as the inferred projected rotational velocities are listed in rotational velocities, may result from contamination by binary Table 2. stars. Since their locations in the CMD coincide with the equal-
5 The Astrophysical Journal, 876:113 (9pp), 2019 May 10 Sun et al.
Figure 5. (Left) CMD of NGC 5822 with the member stars color-coded by their rotational velocities. The best-fitting isochrone is shown as the red curve. A clear trend between stellar rotation and their loci in the CMD region is seen, in the sense that the rapid rotators (yellow) tend to lie on the red side of the eMSTO while the slow rotators (blue) are usually found on the blue side. (Right) Two sample spectra of a slow rotator (top) and a fast rotator (bottom).Hβ and Mg I triplets of the same object are shown in the left- and right-hand columns, respectively. For each spectrum, the best-fitting models are presented as orange curves. mass binary sequence, they are likely unresolved binaries, and NGC 6705 (Marino et al. 2018a) confirmed the existence particularly the star with Gaia ID 5887671397119565312, of slowly and rapidly rotating populations and found that these whose low mass (1.2–1.3 Me) is close to the minimum mass two subgroups are well separated in projected rotational for large stellar rotation. With such a low mass, stars brake velocity, with a difference in mean visin greater than efficiently early on the MS and evolve back to the nonrotating 100 km s-1. Meanwhile, in the intermediate-age clusters NGC tracks (Georgy et al. 2018). Therefore, stellar rotation is 2818 (Bastian et al. 2018) and NGC 5822, such a result is unlikely the cause of such a large shift in color, and binary stars barely seen, which may due to small number statistics and might be a plausible explanation of these two outliers. selection effects. Therefore, we estimated the rotational velocities for all MSTO stars in NGC 5822 based on our synthetic cluster to check the distribution of the stellar rotation 4. Discussion and Conclusions rates. On the basis of previous analyses, we argue that the NGC 5822 is an intermediate-age (0.9 Gyr) Galactic open synthetic cluster can properly reproduce the observed results cluster exhibiting an eMSTO. Through membership determina- and that it can be taken as a reasonable approximation to the tion based on Gaia proper motions and parallaxes, we real cluster. Thus, for each member star in the NGC 5822 investigated the CMDs of NGC 5822 and confirmed that the MSTO region, we inferred the rotational velocities by taking eMSTO is not likely an artifact caused by differential the average velocities of the nearest 50 stars in the extinction. By exploiting SALT/RSS data, we derived the synthetic CMD. projected rotational velocities of 24 member stars and found The distributions of projected rotational velocity visin and v that stellar rotation is strongly correlated with the stellar loci in rot are presented in Figure 7. We found that the projected the CMD in the MSTO region. The red side of the eMSTO is rotational velocities show a dip around 150 km s-1, similar to occupied by fast rotators while the blue side is mainly the results for NGC 1818 and NGC 6705. However, we suggest composed of slow rotators. By comparison with a synthetic that this is an artifact caused by projection effects. In Figure 7 cluster, we have shown that the eMSTO of NGC 5822 can be the “slow” rotators have a peak at 100 km s-1 and they have a properly reproduced and the rotational velocities of the eMSTO dearth of stars with visin~ 50 km s-1, which is different from stars follow the same pattern as that predicted by the stellar the results for the young clusters where the slowly rotating rotation model. populations have lower mean velocities and do not show a gap Combined with NGC 6705 (250 Myr; Marino et al. 2018a) in slowly rotating stars. The distribution of the true rotational and NGC 2818 (800 Myr; Bastian et al. 2018), we have velocities is also shown in Figure 7, and the fact that it shows a confirmed the correlation between stellar rotation and stellar single peak around 200 km s-1 further confirms that the positions in the eMSTO/split MS in three Galactic open equatorial velocities in NGC 5822 should follow a unimodal clusters. Cordoni et al. (2018) found that Galactic open clusters distribution. On the other hand, projection effects are unlikely also follow the trend between cluster age and the extent of to explain the large difference in projected rotational velocities eMSTO seen in Magellanic Cloud clusters (Yang et al. 2013), found in young clusters, and the true rotation rates of MSTO suggesting that they should be regulated by a similar stars in young clusters should in all probability show a bimodal mechanism. Split MSs are believed to be composed of two distribution given the fact that the split MSs can be separated stellar populations characterized by different rotation rates: the into distinct sequences in the CMD. Since the typical masses of blue MS is composed of slow rotators and the red MS is the split MSs in young clusters ( 2.5 Me) and eMSTOs in composed of rapid rotators (Milone et al. 2016). Spectroscopic intermediate-age clusters (1.4 Me–2 Me) are different, if a split surveys of the young clusters NGC 1818 (Marino et al. 2018b) MS and eMSTO are present in the evolutionary sequence of
6 The Astrophysical Journal, 876:113 (9pp), 2019 May 10 Sun et al.
Table 2 Properties of Member Stars with Rotational Velocity Measurements
a a a -1 b -1 c Gaia ID G (mag) Gbp (mag) Grp (mag) T (K) log g Fe/H visin (kms ) vrv (kms ) (1)(2)(3)(4)(5)(6)(7)(8)(9) 5887641469783582208 12.28 12.50 11.92 6900±100 3.5±0.1 0.0±0.5 82.0±12.4 −17.8±5.8 5887666444964220416 11.21 11.38 10.93 7400±100 3.9±0.1 0.0±0.5 182.3±6.1 −30.2±7.0 5887666444964220416d 11.21 11.38 10.93 7200±100 3.6±0.1 0.0±0.5 176.9±6.1 −25.2±5.1 5887669198096568960 12.03 12.27 11.64 6800±100 4.2±0.1 −0.5±0.5 110.8±8.1 −42.0±6.1 5887668648340678784 10.89 11.07 10.61 7200±100 4.4±0.1 0.0±0.5 16.0±38.8 −30.3±6.6 5887642466216028032 11.82 12.02 11.50 7100±100 4.1±0.1 −0.5±0.5 206.2±5.8 −24.1±6.7 5887642466216028032d 11.82 12.02 11.50 7000±100 4.2±0.1 −0.5±0.5 211.3±5.6 −23.1±14.1 5887668575267987840 11.92 12.10 11.61 6900±100 3.7±0.1 −0.5±0.5 163.0±6.1 −26.8±13.6 5887698644394381952 10.65 10.87 10.31 7200±100 3.6±0.1 0.0±0.5 42.7±24.7 −40.1±6.9 5887698644394381952d 10.65 10.87 10.31 7200±100 3.5±0.1 0.0±0.5 53.9±20.3 −33.4±4.4 5887719054073235584 11.07 11.27 10.76 7400±100 4.0±0.1 0.0±0.5 177.6±6.1 −31.6±5.2 5887719054073235584d 11.07 11.27 10.76 7000±100 4.0±0.1 −0.5±0.5 181.7±6.1 −32.9±6.5 5887644935768013056 11.53 11.72 11.21 7400±100 4.1±0.1 0.0±0.5 194.0±6.0 −42.3±5.2 5887644935768013056d 11.53 11.72 11.21 7400±100 4.0±0.1 0.0±0.5 194.2±6.0 −46.4±9.5 5887671255327574144 12.08 12.29 11.75 7000±100 3.5±0.1 0.0±0.5 33.4±29.0 −21.5±6.0 5887695414549752704 10.92 11.13 10.60 7300±100 3.8±0.1 0.0±0.5 120.9±7.3 −36.6±6.6 5887695414549752704d 10.92 11.13 10.60 7200±100 3.7±0.1 0.0±0.5 124.9±7.0 −26.0±6.8 5887718740485998208 11.19 11.43 10.83 7000±100 3.6±0.1 0.0±0.5 246.9±2.4 −27.7±8.6 5887718740485998208d 11.19 11.43 10.83 7100±100 4.3±0.1 −0.5±0.5 239.0±3.5 −33.2±7.7 5887668197310871168 12.04 12.23 11.72 7100±100 3.7±0.1 0.0±0.5 89.0±11.0 15.9±7.1 5887642981611923200 11.66 11.88 11.30 7000±100 4.2±0.1 −0.5±0.5 249.1±2.1 −28.2±6.8 5887642981611923200d 11.66 11.88 11.30 6900±100 3.5±0.1 −0.5±0.5 249.6±2.0 −26.5±9.7 5887722313898780672 12.02 12.22 11.69 7400±100 4.3±0.1 0.0±0.5 120.0±7.3 −31.9±5.2 5887722313898780672d 12.02 12.22 11.69 7300±100 4.1±0.1 0.0±0.5 143.0±6.3 −26.9±8.1 5887671397119565312 12.74 13.04 12.26 6300±100 4.0±0.1 −0.5±0.5 48.2±22.4 −28.6±8.4 5887698296441666688 11.97 12.17 11.64 7400±100 3.9±0.1 0.5±0.5 108.2±8.3 −31.9±5.0 5887698296441666688d 11.97 12.17 11.64 7400±100 3.7±0.1 0.5±0.5 83.7±12.0 −29.3±5.5 5887697987204027264 11.14 11.33 10.85 7400±100 3.8±0.1 0.0±0.5 178.9±6.1 −31.6±8.9 5887697987204027264d 11.14 11.33 10.85 7400±100 3.9±0.1 0.0±0.5 175.1±6.1 −38.0±6.1 5887671534558688640 12.79 13.04 12.38 6500±100 3.7±0.1 −0.5±0.5 138.4±6.4 −21.8±7.8 5887670224535430272 12.26 12.47 11.91 6600±100 3.5±0.1 −0.5±0.5 34.1±28.7 −58.6±8.3 5887671431479631360 12.77 13.02 12.36 6600±100 3.7±0.1 −0.5±0.5 144.0±6.3 −30.3±7.0 5887665895208406272 12.01 12.21 11.66 7000±100 3.9±0.1 −0.5±0.5 201.6±5.9 −32.7±8.1 5887665895208406272d 12.01 12.21 11.66 7000±100 4.4±0.1 −0.5±0.5 202.4±5.9 −42.0±6.7 5887667544475841152 12.14 12.36 11.77 6800±100 4.0±0.1 −0.5±0.5 65.2±16.6 −51.4±4.4 5887642805464232704 12.24 12.45 11.87 7200±100 4.3±0.1 0.5±0.5 86.0±11.6 −14.4±9.7 5887642805464232704d 12.24 12.45 11.87 7000±100 4.3±0.1 0.5±0.5 123.5±7.1 −0.0±13.6
Notes. (1) Gaia DR2 ID; (2), (3), (4) extinction-corrected Gaia bands; (5) effective temperature; (6) surface gravity; (7) metallicity; (8) projected rotational velocity; (9) radial velocity. a Uncertainty adopted from the step size of the model grid. b Uncertainty estimated from the mock test. c Uncertainty given by the MCMC procedure. d Duplicate observation in program 2018-1-SCI-006. star clusters, these two distributions of stellar rotation may We also estimated the stellar masses following similar coexist in the same cluster. This may hint that the stellar procedures as Sun et al. (2018). In essence, we generated a rotation distribution in clusters follow a similar pattern as the synthetic cluster of 10,000 stars with the initial masses field population in the sense that stars more massive than generated through Monte Carlo sampling of a Kroupa stellar ( ) 2.5 Me show a bimodal equatorial velocity distribution while initial mass function Kroupa 2001 . Then, we calculated the G less massive stars have a unimodal rotation distribution (Zorec ratio of the number of member stars with magnitudes from & Royer 2012). However, spin alignment in clusters may play 12.5 to 13.5 mag to that of the synthetic cluster for the same an overlooked role. Corsaro et al. (2017) found evidence of magnitude range. We multiplied the integrated mass obtained for the synthetic cluster by this ratio to estimate the total stellar spin alignment among the red giant stars in the two old open 3 mass in the cluster, 1.7±0.3×10 Me. We confirmed that clusters NGC 6791 and NGC 6819. Lim et al. (2018) inferred changing the magnitude range will not affect the estimation of the v and the inclination angles i of MSTO stars in NGC rot the total mass significantly. Bastian et al. (2018) reported a 6705 from Monte Carlo simulations where v has a linear rot mass of 2800 Me for NGC 2818. These results suggest that distribution and i a Gaussian distribution. They argued that 4 5 eMSTOs are not exclusive to massive clusters (10 –10 Me). cluster members have highly aligned spin axes, which implies a One possible source that could also give rise to a broadened link between stellar rotation and rotational kinetic energy in the MSTO is stellar variability. Salinas et al. (2016) argued that the progenitor molecular cloud. instability strip intersects with the MSTO region of a cluster
7 The Astrophysical Journal, 876:113 (9pp), 2019 May 10 Sun et al.
Figure 6. (Left) CMD of the observed cluster (gray dots), as well as the synthetic cluster, with colors representing the projected rotational velocities. Larger points with white borders reflect measurements of member stars. The adopted ridgeline is shown as a blue dashed line. The black arrow indicates how stellar rotation affects the locus of a star in the CMD. (Middle) Realistic synthetic cluster with a number of stars comparable to that in the observed CMD. (Right) Pseudo-color vs. rotational velocity. Stars with projected rotational velocity measurements are marked as orange dots, and the synthetic cluster stars are marked as gray dots.
11633005, 11473037, and U1631102. R.d.G. is grateful for support from the National Key Research and Development Program of China through grant 2017YFA0402702 from the Chinese Ministry of Science and Technology (MOST). L.D. also acknowledges support from MOST through grant 2013CB834900. Facility: SALT (RSS). Software: PySALT (Crawford et al. 2010), PARSEC (1.2S; Bressan et al. 2012), Astropy (Astropy Collaboration et al. 2013), Matplotlib (Hunter 2007), SYNSPEC (Hubeny & Lanz 1992), emcee (Foreman-Mackey et al. 2013).
ORCID iDs Weijia Sun https://orcid.org/0000-0002-3279-0233 Richard de Grijs https://orcid.org/0000-0002-7203-5996 Licai Deng https://orcid.org/0000-0001-9073-9914 Michael D. Albrow https://orcid.org/0000-0003-3316-4012
References
Astraatmadja, T. L., & Bailer-Jones, C. A. L. 2016, ApJ, 832, 137 Figure 7. Distributions of projected rotational velocities visin (blue) and true Astropy Collaboration, Robitaille, T. P., Tollerud, E. J., et al. 2013, A&A, v ( ) rotational velocities rot orange for member stars in the MSTO region of NGC 558, A33 5822. Both velocities were estimated by taking the average velocities of the Bastian, N., Cabrera-Ziri, I., Niederhofer, F., et al. 2017, MNRAS, 465, 4795 nearest 50 stars in the synthetic CMD. Bastian, N., & de Mink, S. E. 2009, MNRAS, 398, L11 Bastian, N., Kamann, S., Cabrera-Ziri, I., et al. 2018, MNRAS, 480, 3739 with an age of ~13Gyr– and can make a significant Brandt, T. D., & Huang, C. X. 2015a, ApJ, 807, 58 Brandt, T. D., & Huang, C. X. 2015b, ApJ, 807, 25 contribution to the observed eMSTOs. Follow-up observations Brandt, T. D., & Huang, C. X. 2015c, ApJ, 807, 24 of NGC 1846 revealed the presence of a group of (mainly δ Bressan, A., Marigo, P., Girardi, L., et al. 2012, MNRAS, 427, 12 Scuti) variable stars around the eMSTO region. However, the Buckley, D. A. H., Swart, G. P., & Meiring, J. G. 2006, Proc. SPIE, 6267, number fraction of variable stars was not sufficient to produce 62670Z ( ) Cantat-Gaudin, T., Jordi, C., Vallenari, A., et al. 2018, A&A, 618, A93 the observed width of the eMSTO Salinas et al. 2018 . Certain Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245 types of variable or binary stars (e.g., EA-type eclipsing Carraro, G., Anthony-Twarog, B. J., Costa, E., et al. 2011, AJ, 142, 127 binaries) exhibit large changes in radial velocity over time and Claret, A. 2000, A&A, 359, 289 can be detected through multi-epoch observations. However, Cordoni, G., Milone, A. P., Marino, A. F., et al. 2018, ApJ, 869, 139 we did not find such candidates in our sample because of the Correnti, M., Goudfrooij, P., Bellini, A., et al. 2017, MNRAS, 467, 3628 Corsaro, E., Lee, Y.-N., García, R. A., et al. 2017, NatAs, 1, 0064 limited spectral resolution and the small number of observa- Crawford, S. M., Still, M., Schellart, P., et al. 2010, Proc. SPIE, 7737, 773725 tions. Follow-up photometric and spectroscopic observations Dias, W. S., Alessi, B. S., Moitinho, A., et al. 2002, A&A, 389, 871 are required to further investigate the role of variability in Dupree, A. K., Dotter, A., Johnson, C. I., et al. 2017, ApJL, 846, L1 shaping the morphology of the eMSTO region. Espinosa Lara, F., & Rieutord, M. 2011, A&A, 533, A43 Foreman-Mackey, D., Hogg, D. W., Lang, D., et al. 2013, PASP, 125, 306 Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2018, A&A, 616, A1 R.d.G. and L.D. acknowledge research support from the Gaia Collaboration, Prusti, T., de Bruijne, J. H. J., et al. 2016, A&A, 595, A1 National Natural Science Foundation of China through grants Georgy, C., Charbonnel, C., Amard, L., et al. 2018, A&A, 622, A66
8 The Astrophysical Journal, 876:113 (9pp), 2019 May 10 Sun et al.
Georgy, C., Ekström, S., Granada, A., et al. 2013, A&A, 553, A24 Marino, A. F., Milone, A. P., Casagrande, L., et al. 2018a, ApJL, 863, L33 Georgy, C., Granada, A., Ekström, S., et al. 2014, A&A, 566, A21 Marino, A. F., Przybilla, N., Milone, A. P., et al. 2018b, AJ, 156, 116 Goudfrooij, P., Puzia, T. H., Kozhurina-Platais, V., et al. 2011, ApJ, 737, 3 Milone, A. P., Bedin, L. R., Piotto, G., et al. 2009, A&A, 497, 755 Huang, W., & Gies, D. R. 2006, ApJ, 648, 580 Milone, A. P., Bedin, L. R., Piotto, G., et al. 2015, MNRAS, 450, 3750 Huang, W., Gies, D. R., & McSwain, M. V. 2010, ApJ, 722, 605 Milone, A. P., Marino, A. F., D’Antona, F., et al. 2016, MNRAS, 458, 4368 Hubeny, I., & Lanz, T. 1992, A&A, 262, 501 Nataf, D. M., Gould, A., Fouqué, P., et al. 2013, ApJ, 769, 88 Hunter, J. D. 2007, CSE, 9, 90 Niederhofer, F., Georgy, C., Bastian, N., et al. 2015, MNRAS, 453, 2070 Kamann, S., Bastian, N., Husser, T.-O., et al. 2018, MNRAS, 480, 1689 O’Donnell, J. E. 1994, ApJ, 422, 158 Kroupa, P. 2001, MNRAS, 322, 231 Palacios, A., Gebran, M., Josselin, E., et al. 2010, A&A, 516, A13 Kurucz, R. L. 2005, MSAIS, 8, 189 Salinas, R., Pajkos, M. A., Strader, J., et al. 2016, ApJL, 832, L14 Li, C., de Grijs, R., & Deng, L. 2014a, Natur, 516, 367 Salinas, R., Pajkos, M. A., Vivas, A. K., et al. 2018, AJ, 155, 183 Li, C., de Grijs, R., & Deng, L. 2014b, ApJ, 784, 157 Schlafly, E. F., & Finkbeiner, D. P. 2011, ApJ, 737, 103 Lim, B., Rauw, G., Nazé, Y., et al. 2018, NatAs, 3, 76 Sun, W., Li, C., de Grijs, R., et al. 2018, ApJ, 862, 133 Mackey, A. D., & Broby Nielsen, P. 2007, MNRAS, 379, 151 Yang, W., Bi, S., Meng, X., et al. 2013, ApJ, 776, 112 Mackey, A. D., Broby Nielsen, P., Ferguson, A. M. N., et al. 2008, ApJL, Zacharias, N., Urban, S. E., Zacharias, M. I., et al. 2000, AJ, 120, 2131 681, L17 Zorec, J., & Royer, F. 2012, A&A, 537, A120
9